CIB W078 - Information Technologyfor Construction

CIB Publication 361

Proceedings

W078 - Special Track 18th CIB World Building Congress

May 2010 Salford, United Kingdom















CIB
WORKING
COMMISSION
W078
‐
INFORMATION
TECHNOLOGY
IN
CONSTRUCTION







PAPERS
AND
POSTGRADUATE
PAPERS
FROM
THE
SPECIAL
TRACK


HELD
AT
THE
CIB
WORLD
BUILDING
CONGRESS
2010,
10‐13
MAY
2010


THE
LOWRY,
SALFORD
QUAYS,
UNITED
KINGDOM







Selected
papers
from
the
Proceedings
of
the
18th
CIB
World
Building
Congress.
Proceedings
edited
by:
Professor
Peter
Barrett,
Professor
Dilanthi
Amaratunga,
Dr.
Richard


Haigh,
Dr.
Kaushal
Keraminiyage
and
Dr.
Chaminda
Pathirage



W078
Special
Track
Papers
(excluding
Postgraduate
Papers)
reviewed
by:
Dr.
Alain
Zarli,
Dr.
Bill
East,
Mr.
Ersen
Firat,
Dr.
Dana
Vanier,
A/Prof.
Jos
van
Leeuwen,
Ms.
Kathryn
Davies,
Dr.


Yum
Kwok‐Keung,
Dr.
Matt
Prins,
Dr.
Nicola
Maiellaro,
Dr.
Marc
Bourdeau,
Dr.
Marja
Naaranoja,
Prof.
Zhiliang
Ma,
Mr.
Praveer
Kumar,
Dr.
Sami
Kazi,
Dr.
Thomas
Grisham
and


A/Prof.
Robert
Amor






CIB
Publication
361















W078
‐
INFORMATION
TECHNOLOGY
IN
CONSTRUCTION



PAPERS
AND
POSTGRADUATE
PAPERS
FROM
THE
SPECIAL
TRACK

The
objectives
of
the
Working
Commission
are
to
foster,
encourage
and
promote
research
and
development
in
the
application
of
integrated
IT
throughout
the
life‐cycle
of
the
design,
construction
and
occupancy
of
buildings
and
related
facilities,
to
pro‐actively
encourage
the
use
 of
 IT
 in
 Construction
 through
 the
 demonstration
 of
 capabilities
 developed
 in
collaborative
 research
projects
and
 to
organise
 international
 cooperation
 in
 such
activities
and
to
promote
the
communication
of
these
activities
and
their
results.
The
aim
of
W078's
work
 is
 broad
 in
 terms
 of
 the
 design,
 construction
 and
 occupation
 and
 occupancy
 of
constructed
facilities,
but
primarily
it
relates
to
the
integration
and
communication
of
data,
information
and
knowledge
in
the
facility's
life
cycle.


CONTENTS

Papers

A
Building
information
Modelling
Based
Production
Control
System
for
Construction
 1
Sacks,
R.
Radosavljevic,
M.
Barak,
R.

RCM‐Plan:
A
Computer
Prototype
for
Improving
Planning
Reliability
from
a
Lean
 14
Production
Viewpoint
Gonzalez,
V.
Alarcon,
L.F.
Ulloa,
H.

Building
Information
Modelling
in
the
Netherlands;
A
Status
Report
 28
van
Nederveen,
S.
Beheshti,
R.
Willems,
P.

Radio
Frequency
Identification
(RFID)
and
the
Lean
Construction
Process
 41
Taylor,
J.M.

The
Last
Mile:
Best
Practices
for
Successful
Implementation
of
Mobile
Communication
 54
Technologies
at
the
Construction
Site
Lasker,
G.
Cox,
R.F.
Orczyk,
J.J.
Converse,
D.

Real‐Time
Management
in
a
BIM
Model
with
RFID
and
Wireless
Tags
 67
Sattineni,
A.

Empirical
Application
of
GPS
Fleet
Tracking
Technology
to
a
Soil
Excavation
Process
 76
Kang,
J.
Ahn,
S.M.

A
Synopsis
of
the
Handbook
of
Research
in
Building
Information
Modeling
 84
Isikdag,
U.
Underwood,
J.

A
Comperative
Aanalysis
of
the
Strategic
Role
of
ICT
in
UK
and
Turkish
Construction
 97
Industries
Underwood,
J.
Isikdag,
U.
Goulding,
J.
Kuruoglu,
M.

Digital
Image
Processing
for
Evaluating
Defect
Level
in
Visual
Quality
Inspection
 110
Peansupap,
V.
Lorfor,C.

Towards
a
Smart,
Energy‐Efficient
ICT‐Empowered
Built
Environment:
The
REEB
 121
Strategic
Research
Agenda
Zarli,
A.
Bourdeau,
M.
Hannus,
M.
Hassan,
T.

Postgraduate
Papers

Building
Information
Modelling
Processes:
Benefits
for
Construction
Industry
 137
Olatunji,
O.A.
Sher,
W.D.
Gu,N.
Ogunsemi,
D.R.




Impact
of
Information
Technology
to
Facilitating
Communication
and
Collaboration
 152
in
Libyan
Public
Sector
Organisations
Bezweek,
S.
Egbu,
C.

Building
Information
Modelling:
Literature
Review
on
Model
to
Determine
the
Level
of
 168
Uptake
by
Organisation
Haron,
A.T.
Marshell‐Ponting,
A.
Aouad,
G.

Agent‐Based
Negotiation
Mechanism
for
Automatic
Procurement
in
Construction
 185
Ho,
C.
Shih,
H.

Preliminary
Performance
Evaluation
of
an
ORDB‐Based
IFC
Server
and
an
RDB‐Based
 192
IFC
Server
Using
the
Bucky
Benchmark
Method
Jeong,
J.
Lee,
G.
Kang,
H.

A
Conceptual
Framework
for
Research
in
Spatial
Data
Sharing
 201
Salleh,
N.
Khosrowshahi,
F.

The
Changing
Perception
in
the
Artefacts
used
in
the
Design
Practice
through
BIM
 212
Adoption
Coates,
P.
Arayici,
Y.
Koskela,
L.
Usher,
C.


CIB
Brochure
 224

Disclaimer
 226


A Building Information Modelling Based Production Control System for Construction

Technion, S.R.

Israel Institute of Technology

(email: cvsacks@techunix.technion.ac.il)

Radosavljevic, M.

University of Reading

(email: m.radosavljevic@reading.ac.uk)

Technion, B.R.

Israel Institute of Technology

(email: ronen@rbarak.com)

Abstract

We propose visual interfaces for BIM-based construction management systems to empower

construction personnel on site by providing them with construction process and status information.

Computer-aided visualization, not only of the construction product, but also of the construction

process, can provide a unique service to support decision-making by workers, supervisors and

managers, with the goal of achieving stable flows. The three main user interfaces – for a) detailing of

work packages into finer-grained task definitions by the trade managers and preparation of proposed

weekly work plans, for b) collaborative planning and integration between the plans of the different

trade crews, and for c) day to day communication of product and process information to and from

work crews – have been designed and implemented as functional mock-ups. They have been

evaluated in three focus group workshops with project engineers, construction site supervisors, trade

crew leaders and logistics managers. The findings at this early stage are that the interfaces provide

rich information for production control, including monitoring of current process status, and fulfil the

guiding principles defined for BIM-enabled production control. However, significant R&D is still

needed for back-end integration of the various information system components of the system

architecture before the system can be implemented and tested on site.

Keywords: building information modelling, lean construction, production management, visualization

1

mailto:ronen@rbarak.com

1. Introduction

With few notable exceptions, the majority of academic and industrial research on computer-aided

design and visualization in construction has dealt with building design and with pre-construction

planning. There has been far less effort to develop Building Information Modelling (BIM) based tools

to support coherent production management on site. The neglect of production management on the

part of researchers of IT in construction reflects the decline in attention paid to production

management on the part of general contractors and construction managers (Ballard 2000). For various

reasons, construction companies have adopted a business practice of reducing core staff to a

minimum and implementing work through subcontracting (Sacks and Harel 2006). At the same time,

lean construction thinking applied to construction production systems has increased awareness of the

benefits of stable work, of pull flow of teams and materials to reduce inventories of work in progress,

and of process transparency to all involved. 3D visualizations of process status and future direction,

delivered to all on site, are either essential or at least highly beneficial for all of these (Formoso,

Santos et al. 2002). They can empower people working on site to manage the day to day flow of

construction operations with less direct control from higher levels of management, with better quality

and less waste (Sacks, Treckmann et al. 2009).

Production control in construction on site can be facilitated through use of the Last Planner System™

(LPS) (Ballard 2000). In prefabrication projects, methods such as the Process Planning Methodology

(PPM) have proved effective (Radosavljevic and Horner 2007) . Application of the LPS enables trade

managers and construction engineers to collaboratively prepare weekly construction plans that are

feasible and have a reasonably high chance of being executed as planned. The system works by

empowering those who carry direct responsibility for executing work to participate in planning the

work. It is based on the principles of flow defined in lean production texts and the Transformation-

Flow-Value TFV theory (Koskela 2000) of production in construction.

However, the LPS does not overcome all of the difficulties nor does it remove all of the waste

inherent in construction. In practice, percent plan complete (PPC) measures do not reach 100%

(research has shown that the best sites achieve approximately 80% PPC (Bortolazza, Costa et al.

2005)). By definition, lean systems are always subject to continuous improvement (Womack and

Jones 2003), and the LPS is no exception. One of the reasons for this is that construction systems are

uncertain and subject to process change within the time frame of the weekly planning window, so that

filtering tasks for maturity on a weekly level cannot ensure complete process stability. Another is that

the delivery of product and process information to workers can at times be ineffective or inefficient:

product information is provided in the form of drawings and specifications, which contain

inaccuracies or errors. Process information is scant, inaccurate and incomplete: trade crews are

generally uninformed about delays in material deliveries, unavailability of equipment previously

committed to them, or changes in the work plans of crews working in their vicinity. Where they are

informed, it is often too late for them to adapt their own plans. Seppanen (Seppanen 2009) provided

empirical evidence of the systematic failure of traditional production control to manage short-term

decision-making concerning trade crews' progress through a building, resulting in unstable plans and

low productivity.

2

Sacks et al. (Sacks, Radosavljevic et al. 2010) proposed to address these shortcomings by increasing

the degree of resolution for planning and responsive re-planning to a daily level, with the support of a

production planning, control and feedback information system based on building information models.

The system proposed is called ‘KanBIM’, denoting the implementation of a lean production system

with pull flow control (symbolized by the Kanban method) (Ohno 1988) using a building information

modelling (BIM) (Eastman, Teicholz et al. 2008) based information system.

2. Background

A BIM-based lean production management system for construction must enable:

1. visualization of the construction process and its status;

2. visualization of the construction product and work methods;

3. support for planning, negotiation, commitment and status feedback;

4. implementation of pull flow control;

5. maintenance of work flow and plan stability;

6. formalization of production experiments for continuous process improvement.

These principles emphasize the role of a KanBIM system in supporting human decision making,

negotiation among trade crews to coordinate weekly work plans, reduction of the granularity of

planning to a daily level, real-time evaluation of task constraints to compute task maturity, and

implementation of the language/action perspective.

Some BIM solutions have expanded their capabilities by adding 4D functionality. Among these are

'Tekla Structures' (Tekla 2008) and ‘Virtual Construction’ (VICO 2007). Since Tekla's core

functionality is detailing steel and precast concrete structures, its construction management tool

emphasizes fabrication and delivery control based on planned erection dates. As with most 4D BIM

tools, users can link tasks to model objects and use critical path methods to schedule them. To

support fabrication and delivery control Tekla allows users to schedule each piece, within the task,

individually. Virtual Construction’ integrates a BIM model with Location Based Scheduling (LBS),

which uses an underlying CPM network solver. It is based on a bill of quantities and a set of working

rates and costs for resources that are linked to tasks, and enables representation of the schedule as a

line of balance chart as well as in Gantt chart form. These tools, and others like them, include 4D

model visualization but do not support the collaborative production level planning that is essential for

trade managers and crew leaders on site.

Specialized 4D construction planning software, such as 'Synchro Professional' (Synchro 2007),

provide project scheduling, construction visualization, synchronization with design changes, supply

chain management and virtual construction simulation. They do not have internal scheduling

capabilities or an integrated BIM solution but instead allow users to import both schedule and 3D

model from various other applications.

3

A small number of applications have been developed to support Last Planner System™

implementations, but they do not use building models to support visualization. CICLOPS (Evolution-

IP 2009) is an internet application that allows its users to collaboratively prepare and control weekly

work plans. CICLOPS calculates the percent plan complete (PPC) for each weekly plan and can

perform 'Non-Completion Analysis' based on the user's recording of the reasons for tasks being

delayed. ‘WorkPlan’ is a planning tool, developed in research (1999), that applies a database of work

packages and constraints to support work planning. SPS (Koerckel and Ballard 2005) is a commercial

package that helps reduce supply chain variations.

The LEWIS system (Sriprasert and Dawood 2003) represents the most advanced attempt to date to

compile a construction production management system that fulfils the KanBIM principles. However,

it falls short in making the process status visible to work teams on site, it does not explicitly facilitate

negotiation and collaboration in work planning, and it does not implement pull flow control of work.

Thus most 4D solutions incorporate scheduling tools and a connection to a BIM model, but their core

use is for planning and visualizing the process. With the exception of LEWIS, they are not intended

for ‘real time’ production control. They lack the ability to detail work plans with sufficiently fine-

grained resolution, and they have no tools for delivery of information to the work face or reporting

from it, or for assisting real-time decision-making during the course of production itself.

3. Research goal and method

The KanBIM concept encompasses a holistic approach to embedding lean production control

processes through delivery of both process and product information to all project participants,

specifically including workers at the work face on site, using a building information model as its

backbone. Since no comparable systems exist, and no existing software could be adapted for the

purposes of evaluation of the proposed KanBIM system, the research method involved three steps

that were performed in three iterations, with the system being refined and re-evaluated in each cycle:

a) Process analysis and system design;

b) Programming of functional mock-ups of its interfaces;

c) Evaluation of the system in focus group workshop evaluation sessions.

System design began with preparation of a detailed ‘future state’ process flow map of the work flow

envisaged for production planning and day to day production control on construction sites. The

information system required to support the process was then derived, and defined in a system

architecture plan. This step also required selection of the delivery methods (hardware) for each

interface.

Functional mock-ups were prepared for the three main user interfaces. The mock-ups sought to

provide sufficiently complete functionality to thoroughly demonstrate the system’s intended modes of

operation. These user-interfaces cover the stages of a) preparation by trades for weekly work

4

planning meetings, b) negotiation between trade crews prior to and during weekly work planning

meetings with the construction management team, and c) day to day interaction with trade crew

leaders on the job site. Programming of the functional mock-ups served not only the evaluation step,

but was in and of itself a formative activity in testing the assumptions made in defining the work flow

and the system architecture, applying to them a rigor that could not have been achieved otherwise.

The user interfaces were evaluated in three focus group workshops which each involved construction

managers, trade crew managers, and crew leaders, held in the UK and in Finland. The remainder of

this paper describes the first two aspects.

4. The KanBIM planning and control process

The process chart shown in Error! Reference source not found. describes the actors in construction

site production management, the information they each generate, a set of ‘activity scenarios’ in which

information is generated, and the way the information is distributed and recorded in the different

information repositories. The process starts with the creation of a Master Plan. In this stage the users

compile and maintain a set of high-level activities and subordinate work packages, and schedule

them, including trade assignments and buffering. High-level resource levelling must also be done for

major equipment and spaces. This is done using existing construction planning tools.

Stop Task

Prepare Look Ahead Plan

PrepareMaster Plan

21

8

Start

Master Planning

Look Ahead Planning(make ready process)

Weekly Work Planning(compile, coordinate/negotiate and

commit)

Daily Work(commit, execute and report)

Compile and Detail Tasks

3

Coordinate & Commit to Weekly

Work Plan

Start Task

5

Negotiate proposed

plan

Yes

Manage Logistics

Task in Progress

Report Task Completion

No

Construction Planner

Section/zone managers

Construction Planner

Section/zone managers

Trade managers

Trade manager

Trade crew leader

Construction Planner

Section/zone managers

Trade managersTrade crew

leadersHealth & Safety Logistics Manager

Construction Planner

Section/zone managers

Health & Safety

Trade crew leader

Trade crew leader

Logistics Manager

Trade crew leader

Site Engineers Inspector Section/zone

manager

End

Problemsduring

execution

No

Inspect Work and Task

Completion

10

Confirm completion

Yes

Figure 1: Process flow model for the KanBIM system, showing defined activities 1 to 10

5

The next stage is look ahead planning (see Error! Reference source not found.). It consists of

breaking down the high-level activities into smaller, manageable work packages, defining logistic and

engineering constraints in the form of connections between activities and assigning equipment and

materials. The master plan and the look ahead plan are done by managers of the general contractor (or

construction management company) and the principal work package subcontractor managers. Both of

these stages are the same as standard LPS, with only one additional requirement, which is that they

are prepared using a BIM interface in which building elements are associated with the activities. This

capability, available in the existing commercial software described above, allows integration of the

product model with the high-level process model. Since 4D functionality is increasingly common in

BIM tools, we assume that an integrated product and process model can be prepared and consider it

as the starting point for the next stages.

The next step of the LPS process, weekly work planning, is divided here into two stages. First, in

activity 3 in Error! Reference source not found., each trade crew details its work packages into a

set of candidate tasks that it can perform during the following week, in preparation for the weekly

work planning meeting (stage 4 in Error! Reference source not found.). This activity starts with a

set of candidate work packages that were drawn from the look-ahead plan according to their planned

start date and priority. Each work package contains a set of 'task types' representing the different

kinds of work needed to perform it according to the production method. For example, in order to

erect a drywall, the following tasks are needed: build the wall frame; close the first side with plaster

boards; place insulation materials and fix any mechanical, electrical or plumbing (MEP) embeds;

close the second side with boards; and apply joint strips, sand and paint. BIM objects can require one

or more task types and the associations are recorded with the object's properties.

The work packages are shown using symbols and highlighted object groups in the model. The trade

contractor’s manager and his or her crew leaders divide the work packages into candidate tasks by

selecting a subset of building elements from the work package elements and grouping them into

distinct tasks according to their task types. For easier selection and better control of the overall

process of dividing the work packages into tasks, all building elements that have not yet been

allocated to tasks are labelled 'unassigned' and highlighted appropriately. The user interface to

support this activity is shown in Figure 2, which shows a hierarchical work package/task type/task

tree, a view of the building model focused on the work package zone and elements with symbols

representing its tasks, and a weekly schedule planning area at the bottom of the screen. Tasks are

scheduled and assigned to crews by dragging their symbols to the rows of specific crews on specific

days. In addition to tasks created and assigned by the trade manager, there are also two kinds of

special tasks: tasks that the trade manager assigns to other supporting trades and tasks that are

assigned to this trade subcontractor by other trades. These tasks need to be assigned to crews in order

to become part of the weekly work plan, in a negotiated process that is explained below.

Since each trade contractor creates its own proposed weekly work plan, the plans need to be

synchronized and finalized to form a mutually agreed project-wide work plan. This is done in a

weekly work planning meeting (activity 4 in Error! Reference source not found.) that is directed by

the project planner and in which all trade managers participate. During the meeting the project

planner reviews the candidate work packages and tasks for promotion to approved tasks for the

6

coming week. The interface for this activity is presented on two large screens: a data view (shown in

Figure 3) and a corresponding model view. The two screens are merely different representations of

the same content (one alphanumerical and the other graphical) and any operation on one, is

automatically reflected in the other. For example, when a task is selected in the data view the model

view will focus on and highlight its building elements and show temporary equipment; or when the

tasks are filtered in one view (by date, space, contractor etc.) the other will show the same results.

The interface allows the users to switch between four different aggregation data views (tasks sorted

by contractors, work packages, spaces and shared equipment) to eliminate any clashes and to improve

plan reliability.

Figure 2: User interface for detailing work packages to tasks and compiling the weekly work plan by

allocating crews to tasks

Any conflicts identified must be resolved through discussion and coordination between the relevant

trade managers. To resolve conflicts they can change their proposed plans using the same interface

used for initial planning (Figure 2). Changes could include rescheduling tasks, assigning more crews

or workers, changing resources by changing construction methods, and others. The changes are made

while all the participants are online so that the project planner views and all 3D model views will

reflect the changed overall weekly work plan.

After applying changes to the plan to make it feasible and acceptable for all the ‘last planners’, each

of them must explicitly accept their part of the plan and commit to executing theirs tasks. Plan

acceptance is shown on the project planner interface and only when a group consensus is achieved is

the weekly plan approved as a whole.

7

The next level of planning takes place on a daily basis, concurrently with execution of the work

through each week. This is the heart of the KanBIM process, where the crew leaders are given direct

access to the work plan and empowered to coordinate their work with all other crews as and when

needed (activities 5, 6 and 8 in Figure 1). The specialized model interface (shown in Figure 4Error!

Reference source not found.), which shows each crew leader’s specific tasks, is delivered via a

large scale touch screen (see Figure 5). This interface not only delivers process and product

information on demand, it also collects process information in real-time. Crew leaders use it to report

the start of tasks as they are begun, to update ongoing tasks according to actual performance, to report

that they have stopped work on a task and report the problem that caused the stoppage, and to report

completion of finished tasks.

Figure 3: Project planner contractor view interface for creating integrated and synchronized weekly

work plans

Problems that adversely affect execution, such as unavailable equipment, can be reported together

with details that enable responders to resolve them, such as details of which specific piece of

equipment is malfunctioning or missing, as shown in Figure 6. In this way crew leaders can also

report design issues directly on the model by using graphic annotation tools and voice messages. The

production management server can alert those responsible for solving the issue according to a

predefined work flow and create action items for fixing it. In the event that a crew leader needs to

change the execution sequence of his/her tasks, they can use this screen to initiate dialog to negotiate

the changes with the project planner and any other relevant crew leaders, in order to maintain overall

plan stability. Any changes are immediately reflected in all model views, so that all project

participants are aware of actual current status.

8

For learning purposes and to improve project performance, when a task is reported complete crew

leaders are asked to report any difficulties even if the task was completed as planned. By pressing the

complete button, the crew leader is also pulling an inspector to approve the completion of the task

(activity 10 in Figure 1). If the task completion is rejected, the rework needs to be re-scheduled by the

project planner and the trade manager.

Figure 4: Trade crew leader work status and reporting interface showing a crew's tasks. The crew

leader can ask the system to show neighbouring tasks for a complete picture of the overall work

Figure 5: Work face specialized model

interface on a 40” touch-screen mounted on a

mobile trolley. The system identifies crew

leaders (by RFID reader or by entry of a

unique ID code) and delivers specifically

tailored information concerning their tasks.

Figure 6: Reporting form for problems during

execution which led to stopping a task. The

reporting tool enables information flow from

the work face to the information servers to

update the work status and to raise flags when

problems are encountered.

The information for each task is organized in a 'control card' according to seven pre-conditions and

constraints: preceding activities, workspace, information (designs and specifications), safety,

9

materials, equipment and crew. For each pre-condition an independent maturity index (MI) is

calculated based on the constraints release status, so that a user can 'see' the maturity status of any

given task. Full details are provided in (Sacks, Radosavljevic et al. 2010).

5. System architecture

Figure 7 provides a high-level view of the system architecture. The main database contains the

construction model, which is a combination of the product model, the process model and the status

model. At the start of any project, they are generated from the design and fabrication models by

applying construction methods (recipes), work package aggregation and compiling temporary

equipment process related objects. Subsequently, the construction BIM modeller is responsible for

synchronization of the database with the design and fabrication models. Interaction between the

KanBIM users and the construction model is facilitated by user interfaces such as look ahead

planning (based on 4D capabilities), weekly plan preparation (Figure 2), weekly work planning and

negotiation (Figure 3), crew leaders' interface for delivering information and reporting status (Error!

Reference source not found.) and an alert system to support organizational work flow. All of these

are based on lean construction processes.

Two separate modules work in the background. The first module generates tasks constraints as soon

as tasks are created. As most constraints are predefined at a higher level for work packages, this

module details the constraints at the task level. The second module computes the maturity index (MI)

and a pull flow index (PFI) for each task. The PFI defines the priority to be assigned to a candidate

task according to the need for that task as determined by the maturities of its successor tasks, which

reflect the downstream demand, or pull.

The sources of the information the system uses extend beyond the boundaries of the construction

product and process mode. Information may reside in different peripheral construction management

systems, such as logistics, purchasing, human resources and personnel control, design management

systems, fabrication management systems and external databases. Sophisticated information or

objects brokers are needed to integrate this information.

10

Figure 7: System architecture schema

6. Conclusions

The principles for development of a KanBIM system have been classified in seven main areas (Sacks,

Radosavljevic et al. 2010): process visualization; product and method visualization; computation and

display of work package and task maturity; support for planning, negotiation, commitment and status

feedback; implement pull flow control; maintain work flow and plan stability and formalize

experimentation for continuous improvement. Some of the ways in which the KanBIM system, as

specified to date, fulfils these principles, are discussed below.

During plan execution, current status visualization is attained using the set of graphical symbols

shown in Figure 4. The symbols describe the current task status: ready, not ready, task in progress,

task stopped, etc. Symbols that represent deviation from plan are supplemented with additional

information, such as maturity level or partial completion indicator.

The BIM is the foundation of the KanBIM system database. A 3D model view serves as a background

platform in all interfaces for conveying project data and navigating through it. The challenge is to

make product and process information ubiquitous at the workface without encumbering crew leaders

or workers with hardware that may hamper their comfort, safety or productivity. This can be achieved

using personal digital assistants, mobile phones or other portable wireless devices, but these all have

limitations, particularly with regard to screen size. The primary solution suggested for implementing

KanBIM interfaces is to use large format all-weather touch-screen monitors which do not impose

physical restrictions on workers, enable discussion among crews who can all view the same model or

animation together, and provide the essential function of easy-to-operate online feedback. This format

also enables easy navigation and data access.

11

The KanBIM system deals with plan stability on two levels: the planning process and the execution.

In planning, it uses the maturity index as the main parameter for deciding which work package or task

will be done during the week. In task detailing (stage 3 in Figure 1), a task can initially be assigned to

a weekly plan even if its maturity is not yet 100%, but this implies a commitment on the part of the

trade manager to release all constraints by the planned execution date. During execution, the KanBIM

system works to maintain plan stability by applying the principle of ‘sticking to plan’ while at the

same time enabling rapid negotiation and thorough coordination of any necessary changes to the plan.

The pitfalls of potential negative impacts on other trades and the danger of ‘making-do’ and

subsequent rework mean that plan changes must be negotiated and recorded. The system enables

negotiation by facilitating ad-hoc toolbox meetings within a crew with real-time information, or

conversations between all those who might be influenced from rescheduling the task so that the new

plan will not compromise their work.

References

Ballard, G. (2000). The Last Planner™ System of Production Control. School of Civil Engineering.

Birmingham, U.K., The University of Birmingham: 192.

Bortolazza, R. C., D. B. Costa, et al. (2005). A quantitative analysis of the implementation of the Last

Planner System in Brazil. 13th Conference of the International Group for Lean Construction. R.

Kenley. Sydney, Australia, UNSW: 413-420.

Choo, H. J., I. D. Tommelein, et al. (1999). "WorkPlan: Constraint-Based Database for Work

Package Scheduling." Journal of Construction Engineering and Management 125(3): 151-160.

Eastman, C. M., P. Teicholz, et al. (2008). BIM Handbook: A Guide to Building Information

Modeling for Owners, Managers, Architects, Engineers, Contractors, and Fabricators. Hoboken, NJ,

John Wiley and Sons.

Evolution-IP. (2009). "CiCLOPS." from http://www.ciclops.info/.

Formoso, C. T., A. D. Santos, et al. (2002). "An Exploratory Study On The Applicability Of Process

Transparency In Construction Sites." Journal of Construction Research 3(1): 35 - 54.

Koerckel, A. and G. Ballard (2005). Return on investment in construction innovation - a lean

construction case study. 13th Annual Conference on Lean Construction, Sydney, Australia, UNSW.

Koskela, L. (2000). An exploration towards a production theory and its application to construction.

VTT Building Technology. Espoo, Helsinki University of Technology: 296.

Ohno, T. (1988). Toyota production system: beyond large-scale production. Cambridge, Mass.,

Productivity Press.

12

http://www.ciclops.info/

Radosavljevic, M. and R. M. W. Horner (2007). "Process planning methodology: dynamic short-term

planning for off-site construction in Slovenia." Construction Management and Economics 25(2): 143-

156.

Sacks, R. and M. Harel (2006). "An economic game theory model of subcontractor resource

allocation behavior." Construction Management & Economics 24(8): 869-881.

Sacks, R., M. Radosavljevic, et al. (2010). "Requirements for Building Information Modeling based

Lean Production Management Systems for Construction." Automation in Construction in press.

Sacks, R., M. Treckmann, et al. (2009). "Visualization of Work Flow to Support Lean Construction."

Journal of Construction Engineering and Management 135(12): 1307-1315.

Seppanen, O. (2009). Empirical Research on the Success of Production Control in Building

Construction Projects. Faculty of Engineering and Architecture. Espoo, Finland, Helsinki University

of Technology PhD: 325.

Sriprasert, E. and N. Dawood (2003). "Multi-Constraint Information Management and Visualisation

for Collaborative Planning and Control in Construction." ITcon - IT in Construction 8: 341 - 366.

Synchro. (2007). "Synchro." from www.synchroltd.com.

Tekla. (2008). "Construction Management." from www.tekla.com.

VICO. (2007). "Virtual Construction 2007." Retrieved April 2007, from

http://www.vicosoftware.com/.

Womack, J. P. and D. T. Jones (2003). Lean Thinking: Banish Waste and Create Wealth in Your

Corporation. New York, Simon & Schuster.

13

http://www.synchroltd.com/http://www.tekla.com/http://www.vicosoftware.com/

RCM-Plan: A Computer Prototype for Improving Planning Reliability from a Lean Production Viewpoint

González, V.

Universidad de Valparaíso

(email: vagonzag@uc.cl)

Alarcón, L.F.

Pontificia Universidad Católica de Chile

(email: lalarcon@ing.puc.cl)

Ulloa, H.

Pontificia Universidad Católica de Chile

(email: hulloa@ing.puc.cl)

Abstract

The creation of reliable work plans level is perhaps one of the most relevant stages in construction

planning process. However, current practice and research in construction are characterized by

developing this process in an informal fashion in which decisions are based mainly on intuition and

experience of project personnel. This paper introduces RCM-Plan, a computer prototype designed to

support operational planning using the so-called Reliable Commitment Model (RCM). RCM is a

statistical modelling approach based on lean production principles, which produces reliable

predictions of work plans, capacity and other production variables for short term-periods using

common field information such as workers, buffers, and planned progress. Thus, RCM promotes a

more reliable operational planning process and improved performance of work plans. The core

functions of RCM-Plan are designed to systematize and automate the procedures associated with the

RCM methodology, generating the analysis and information in a quick, easy way and it is

instrumental to facilitate the application of the RCM concepts. The capabilities of RCM-Plan are

demonstrated using a case study. In addition, limitations and further research related to the RCM-

Plan are addressed.

Keywords: computer prototype, lean production, operational planning reliability, RCM-plan, reliable

commitment model

14

mailto:vagonzag@uc.clmailto:lalarcon@ing.puc.cl

1. Introduction

The creation of reliable work plans has received increased attention as a critical issue in construction

projects performance during the last decades (Ballard, 2000; Tommelein et al, 1999; among others).

The recognition that these plans at the operational level characterize the materialization of a project,

namely its production through construction processes, highlights their decisive role for

accomplishment of project goals. In practice, a reliable work plan is characterized by a high

fulfillment of planned work. However, the decision frame to make and predict work plans has

several limitations in traditional management practices (González et al, 2009). Commonly,

construction projects outsource most of the work to subcontractors, and work plans are arranged

between contractors and subcontractors. Contractors should strive to obtain reliable work plans from

the subcontractors. Thus, many of them assign work to subcontractors based on their intuition and

experience, resulting in unreliable work plans (Sacks and Harel, 2006). Then, the more unreliable

work plans the lower is project performance (Ballard, 2000).

Lean production is a management philosophy focused on adding value from raw materials to finished

product. It allows avoiding, eliminating and/or decreasing waste such as waiting/idles times, rework,

overproduction, excessive movement, among others, from the value stream. One of the main goals of

lean production is reducing variability (Womack and Jones 1996). In construction, variability depicts

varying production rates, labor productivity, schedule control, cost control, etc., which is a well-

known problem due to its detrimental impacts in project performance on which there is much ongoing

research (Ballard, 1993; Tommelein et al, 1999; among others). The Last Planner System (LPS™) is

a production planning and control system based on lean production principles that was developed to

improve planning reliability in construction projects (Ballard, 2000). LPSTM

provides the means and

tools to deal with variability in projects providing a stable production environment and reducing the

negative impacts of variability. This helps create reliable work plans for short-term periods

(operational level). However, last planners still create and predict work plans to make their planning

commitments using mainly their intuition and experience, resulting sometimes in unreliable

commitments (González et al, 2009).

Thus, this paper proposes a computer prototype termed as RCM-Plan to help planners to make more

reliable performance predictions based on data obtained on site. RCM-Plan is based on the Reliable

Commitment Model (RCM) (Gonzalez et al 2010), a lean tool which enhances the creation process of

work plans in repetitive projects. RCM uses statistical models to predict performance and produce

more reliable work plans, using information about workers, buffers, actual and planned progress

avoiding current decision patterns. The core functions of RCM-Plan are designed to systematize and

automate the RCM procedures, generating analyses and information in a quick and easy way. In this

paper, planning reliability improvement using the LPS™, the RCM conceptual and analytical

framework and the computer structure of the RCM-Plan are addressed. Then, a case study is used to

show the use of the RCM-Plan to support the development of reliable work plans.

15

2. Improving planning reliability

The Last Planner System (LPS) is a production planning tool based on lean principles widely used

today in construction. LPS™ acts over four project planning levels: (i) „Initial planning or master

plan‟ (strategic level), which produces the initial project budget and schedule, and provides a co-

ordinating map that „pushes‟ completions and deliveries onto the project. (ii) „Phase Planning‟ a

subdivision of the master plan which is transformed in a “pull plan,” with participation of key

members of the project team during that phase. (iii) „Look ahead planning‟ (breakout of master plan

– tactical level), which details and adjusts budgets and schedules „pulling‟ resources into play. (iv)

„Commitment planning or work plans‟ (short-term period – operational level), which regards the

activities and schedule work that will be done on-site according to the status of resources and

prerequisites (Ballard, 2000).

The traditional management approach for work plans defines activities and schedule work that will be

done, in terms of what should be done from a master plan, compromising crews with no real

consideration for what they are actually able to do. Then, the crew ability to reliably perform work

depends on the stability of the so-called workflow. In construction, workflow can be characterized

by crews moving from location to location and completing the work that is prerequisite to starting

work by the following crew. In turn, a stable workflow depends on what construction preconditions

such as resource (design, components and materials, workers, equipment, space) and prerequisites

(complete work of upstream activities) should be available whenever they are needed. However,

variability of workflow could negatively affect crews‟ performance, causing idle time or ineffective

work (Tommelein et al, 1999).

In contrast, LPS™ provides a predictable production environment in projects, decreasing workflow

variability and creating reliable work plans to derive maximum project benefits. The overarching

criterion in the LPS™ is that activities should only be committed if they can be performed (i.e. all

resources and prerequisites that are needed must be available), transforming what should be done into

what can be done, from which a work plan can be formulated. Thus, work plans will be based on

achievable assignments serving as a commitment to what will actually be done. In this paper, the

notion of reliability is focused on project planning, so that Percentage of Plan Completed (PPC)

depicts a project planning reliability index. PPC is understood as the ratio between actual completed

activities and planned activities in a short term period (typically one workweek). A low PPC means

unreliable planning and a high PPC close to 100%, means the opposite.

LPS™ has been applied in numerous projects around the world in the last fifteen years, and a wide

range of performance improvements have been reported (Alarcon et al 2005, Liu and Ballard 2008,

González et al, 2008a). The main system assumption is that an increase in planning reliability,

measured through PPC, should improve project performance. Recently, several researchers have

demonstrated a positive and strong relationship between planning reliability and project performance

(González et al, 2008a). However, current planning practices at the operational level reduce the

ability to achieve reliable commitments to improve project performance. The RCM model is

designed to support the planners in making more reliable commitments in the short term planning

16

process and it can be used within the LPS process or as an individual method when the LPS is not in

place.

3. Reliable Commitment Model (RCM) framework

RCM provides an operational decision-making tool for predicting work progress in projects using

statistical models. These models use historical information of several production variables such as

labor, Bf, and planned progress, to attain a more reliable planning process at the operational level.

Since RCM is based on lean principles, it helps to reduce production variability by improving

planning reliability and matching work load with labor capacity (more details in González, 2010). In

practice, RCM uses multiple linear regression (MLR) to formulate the model, which assumes the

following form: (González, 2010):

PPWIPBfW PRP 3210 (1)

where:

PRP is the predicted progress for a process in a short-term period (typically one workweek). Units

may be m2, m

3, linear-meters, houses, apartments, etc.

W is the number of workers for a process in a short-term period. W is the sum of workers in the

planning horizon. For instance, if the short-term period is one workweek of 5 days, and there are 5

worker-days, W is 25 worker-weeks.

WIP Bf is the available Work-In-Process Buffer for a process at the beginning of a short-term period.

In general terms, a Buffer allows isolating a production process from the environment as well as the

processes depending on it. Buffers can prevent the loss of throughput, wasted capacity, inflated cycle

times, larger inventory levels, long lead times, and poor customer service by shielding a production

system against variability (more details in Hopp and Spearman, 2000). For instance, for a one

workweek the WIP Bf for the painting process, which depends on the wall-stucco process, is the

available work produced by the wall-stucco process, measured at the beginning of the week, before

painting begins. Units may be m2, m

3, linear-meters, houses, apartments, etc.

PP is the planned progress for a process in a short-term period. Units may be m2, m

3, linear-meters,

houses, apartments, etc.

Only significant variables are selected in the RCM models, since including redundant variables may

lead to incorrect analysis of scenarios. Thus, MLR models with the least number of variables and

with the highest coefficient of determination (R2) are selected. Bustamante (2007) demonstrated that

good quality models are obtained using this heuristic. Furthermore, the prediction accuracy of RCM

is evaluated using two indicators: Process Reliability Index (PRI) and Predicted/Planned

Commitment Confident Level (CCL). PRI measures the degree of process effectiveness from a

commitment standpoint, expressed as:

(2) 100PP

AP PRI

ji,

ji,

17

where:

PRIi,j is the process reliability index for week i and process j (%), i=1…n; j=1...m.

APi,j is the actual progress for week i and process j, i=1…n; j=1...m.

PPi,j is the planned progress for week i and process j, i=1…n; j=1...m.

PRI ranges between 0 and 100%. Note, when AP is higher than PP, PRI is limited to 100%. A low

PRI means unreliable process planning and a PRI close to 100% means the opposite. Meanwhile

Predicted/Planned CCL is a measure of the planning prediction reliability for a process made by both

decision-makers (planned progress) and/or RCM models (predicted progress) and it is expressed as:

(3)

where:

Predicted/Planned CCLi,j is the commitment confidence level for week i and process j (%) for both

predicted and planned PRI.

Predicted PRIi,j is the predicted process reliability index for week i and process j. Predicted PRI

replaces AP by PRP in Eq. (2).

Planned PRIi,j is the planned process reliability index for week i and process j. Its value is estimated

by a decision-maker given a planned progress according to his or her experience.

Actual PRIi,j is the actual or real process reliability index for week i and process j. Actual PRI is

computed using Eq. (2).

Predicted CCL measures the process commitment accuracy for the predicted progress comparing the

predicted and actual PRI. Similarly, Planned CCL compares the planned and actual PRI. When the

ratio in Eq. (3) is less than 0, its value is set to 0.

RCM originally was implemented through nomographs, which relate mathematical and graphically

planned progress with the other production variables (González et al, 2010). Then, since PRI is given

by:

PPPRIPRPPP

PRPPRI (4)

And, if equation (1) is replaced in equation (4), PP can be expressed as:

3

210

PRI

WIPBfWPP (5)

Eq. (5) establishes a relationship between PP and W, WIPBf, and PRI. PP can be either planned by

decision makers or estimated by the RCM. Similarly, PRI can be either planned or predicted. Fig. 1

illustrates a nomograph for a repetitive housing project, showing the interaction between the different

variables involved. Nomographs are commonly used in engineering disciplines (e.g., hydrologic

engineering) and can be easily applied by construction decision-makers, such as project managers,

which can use it to plan activity progress for a given resource frame. RCM methodology to estimate

100ActualPRI

ActualPRIPlannedPRI/edictedPr1PlannedCCL/edictedPr

j,i

j,ij,i

j,i

18

work plans is summarized in Fig. 2. The conceptual and mathematical RCM framework, as well as

the implementation methodology, were comprehensively tested and validated in 9 projects

(multifamily residential, multi-story building and industrial), 23 activities, and a total of 260

workweeks by González et al (2010).

Figure 1: Illustration of a general RCM nomograph to calculate planned progress for next short-term

period, given a pre-determined MLR model (with the assumption that it is a reliable model such that

R2 ≥0.6 and P-value ≤ α=0.05) and different PRI values (adapted from González et al, 2009)

1. Selection of activities

2. Initial data collection for each activity in week i: PP, planned PRI, planned W

WIPBf, AP and actual W.

i) Is the measured horizon plan higher than two weeks?

3. Selection of the best MLR model using the number of variables, R2 and P-values

for week i+2

No

Yes

ii) Is it possible to determine a valid

regression model?No

Yes

4. Def inition of Nomographs for week i+2

5. Def inition of a base case for week i+2 given initial data in which a predicted PRI

value is def ined.

10.Evaluation of the RCM predictions accuracy for week i+2. Indicators:

i)Predicted and Actual PRI, and ii) Predicted

Start

v) Does the RCM process

continue?Finish

No

Yes

Note: To construct a multiple linear regression model is necessary at least two data sets or points. Then, it is the reason to request two planning weeks measured in the f irst decision block.

9.Final data collection for each activity in week i+2: AP and actual Wiii) Is predicted

PRI

sections describe previous tools for improving planning reliability and production predictions in

construction. Then, the RCM-Plan software architecture is introduced.

4.1 Computer approaches to improve planning reliability and production predictions in construction

LPS™ application has inspired the development of several computers applications, supporting

several of its main functions, which allow improving planning reliability. Among others, can be

mentioned: WorkPlan (Choo et al, 1998); WorkMovePlan (Choo and Tommelein, 2000); Integrated

Production Scheduler (IPS) (Chua et al, 1999); Lean Enterprise Web-based System for Construction

(LEWIS) (Sriprasert and Dawood, 2002); SPS Production Manager (SPS|PM) (SPS|PM, 2009);

Project Plus Control (P+C) (P+C, 2009). In general, these approaches provide support to the

planning process but do not solve the operational planning issue that emerges when contractors and

subcontractors arrange work plans making predictions about their performance. In general, many

contractors assign work to subcontractors based on their intuition and experience resulting in

unreliable work plans.

On the other hand, there are several methods in construction that can improve production predictions

such as virtual prototyping (Huang et al, 2007); or discrete event simulation modeling (Martínez,

1996). However, these predictive approaches can be complex to adopt among construction

practitioners as an ordinary practice. RCM-Plan, described in the next section, provides a tool that is

easy to use and implement to automate the procedures associated with RCM, providing prediction

capabilities to support the adoption of reliable commitments in planning meetings.

4.2 RCM-plan framework

RCM-Plan was developed as part of a research effort from the Production Management Center

(GEPUC) at Pontificia Universidad Católica de Chile, to improve the current planning practices in

the Chilean construction industry. RCM-Plan architecture is based on Microsoft Visual Basic .Net,

framework 2.0, developed with Visual Basic 2005. Microsoft Excel version 2000 is required for

RCM-Plan operation on the client-machine user application. The matrix approach is used in RCM-

Plan to calculate the different parameters of the MLR models, i.e. β0…β3, which is implemented in

Excel and in turn interacts with the .Net framework, showing the main MLR statistical results. By

using Excel the system generates the necessary graphics such as nomographs, weekly progress, PRI

evolution, labor resource evolution, etc. and several types of reports. RCM-Plan does not use data

base provided by external suppliers, since its arquitecture uses .xml files to store historical data,

making the information very portable. By summarizing RCM methodology shown in Fig. 2, RCM-

Plan computer architecture and operation is briefly described as follows:

1) Selection /Creation of Activities: In this stage, project managers select activities to improve their

planning reliability. Thus, activities are created within RCM-Plan and are individually analyzed to

20

get their different production outputs and statistical parameters. For one or more activities, RCM-

Plan generates data base files termed as „.rcmpl‟ that can be imported or exported to other computers.

2) MLR Models and Nomograph Construction: The RCM process collects on-site information

and predicts work plans performance on a weekly basis, generating typically MLR models starting

from the 3rd

week. Figure 3 shows a general screen for tools and information that provides RCM-

Plan, listing a specific activity (in this case, partition-joint tape). The main tools are related to Enter

Planning (for actual week), Modify Planning of Last Week, Modify Selected Week (data of any week),

PRI and Progress Graphics (PRI and Progress evolution) and Average PRI (summary of PRIs and

CCLs). In this version, only the predicted CCL is regarded. Also, RCM-Plan considers a predicted

PRI equal to 100% to develop nomographs (a common reliability level expected by planners). In

Figure 3, 17 weeks of historical information are considered and the planning process of 18th week is

illustrated.

Partition Joint-Tape (ML)

Workers (W)

Workers (W)

Workers (W)

Figure 3: General screen of RCM-Plan

To plan 18th week, it is necessary to select a MLR model. RCM model automatically calculates all the

possible combination of variables (W, WIPBf and PP) and its corresponding R2. In Figure 4, the

Adding Planning for Week 18 dialog box, Selection of Function tab, shows which variables are

selected for a specific MLR model and the corresponding R2-value (decision-makers manually chose

a MLR model using the heuristic mentioned earlier in section 3). This dialog box allows showing the

full MLR model (View Prediction Function option) which is a function depending on W and WIPBf

variables in this case and creating the corresponding nomographs (Nomograph option). The latter

option allows showing graphically the interactions between PP and W given several WIPBf sizes.

This is the most important feature of the RCM-Plan, to explicitly visualize in a multidimensional

environment the interactions of several production parameters, allowing reliable prediction of activity

progress, and thus, more reliable planning commitments and work plans.

3) RCM Planning Process: Once a decision-maker has selected PP, W and WIPBf levels, this

information should be entered into the planning for next week (18th week in Figure 5), which is done

in the Adding Planning Week 18 dialog box, Enter Planning Data tab. Note that W is entered as a

planned estimate. Predict Progress is automatically generated as RCM output, showing the PRP level

21

for week 18, i.e. predicted work plans for this week. At the end of the 18th week, AP and actual W

levels are entered using the Modifying Progress dialog box.

Figure 4: Adding planning by creating RCM prediction function and nomographs

Figure 5: Adding planning by entering RCM inputs and actual progress

4) Evaluation and Feedback: At the end of every week, several reports are produced (see Figure 6).

5000 (ML)

3000 (ML)

8000 (ML)

2800

2200

Redefine Chart Limits Print Chart

Worker-Weeks (W)

Addi ng Planning Week 18

Pla

nned

Pro

gres

s (P

P)

MLWorker-Weeks

Actual Progress

Enter progresses and Worker-

Weeks, corresponding to the week 18

Planned progress for this week: 2500

(ML), with 20 planned Worker-Weeks

22

Figure 6: PRI and progress evolution reports

The PRI and Progress Graphics dialog box describes the Predicted/Actual PRI and

Planned/Predicted/Actual Progress respectively for all weeks. Also the Summary of PRI and

Predicted CCL dialog box described the average Predicted/Actual PRI and the average Predicted

CCL for the analysis period. This information allows evaluating performance of planning process in

terms of its reliability and accuracy using the RCM-Plan.

5. Case study

From the total cases studied with the RCM (section 3), RCM-Plan was actively implemented in 6

projects and 15 activities. The RCM-Plan arquitecture allows the automation of RCM calculations

and procedures and a reliable visualization of its mains inputs. These features facilitate the on-site

application of RCM concepts and reduce possible barriers of implementation. In this section, a case

study where RCM-Plan was applied is used to illustrate its impacts on planning reliability and

activity performance.

5.1 Case study A: Multi-family residential building

Plastering activity in a multi-family residential building was selected, analyzing how improving

planning reliability through the RCM-Plan could increase labor productivity. Figure 7 summarizes

the data evolution of RCM-Plan application.

PRI and Progresses Graphics

Select number of weeks to view Print Graphics

Week

Week

Actual PRIPredicted PRI

Planned ProgresPredicted Progres

Actual ProgresWe

ek

%P

lan

nin

g R

elia

bili

ty

23

No-predictions period

0

200

400

600

800

1000

1200

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Pro

gre

ss

(lm

)

Week Planned ProgressPredicted ProgressActual Progress

+∆ PRI 41.0% +∆ productivity 10.3%

Predictions/no-decisions period

Predictions/decisions period

Mean Labor Productivity 20.4 (m2/worker-weeks)

Mean Labor Productivity 22.5 (m2/worker-weeks)

No-predictions period

Figure 7: RCM application evolution case study A (González et al, 2009)

In this case, MLR models were mainly specified as a combination of W and/or WIPBf variables, with

different model parameters, β1 and β2, every week (note that RCM process is dynamic, changing

parameters and even variables of MLR models week to week). Three different periods can be

distinguished in Figure 7: the No-predictions period (1st

– 2nd

weeks), where data is collected for the

RCM; the Predictions/no-decisions period (3rd

– 13th weeks), where planning predictions were

performed to test the RCM, but were not used by the manager to make decisions; and the

Predictions/decisions period (14th -17

th weeks), where the manager relied on the RCM outputs to

make planning decisions.

RCM-Plan was used in the planning process of this case from the beginning, being the RCM-Plan

impact analysis focuses on the last two periods. An active intervention on the W and WIPBf

variables was decided on the 11th week, starting from the 14

th week, when sensitivity analyses using

the RCM-Plan were performed to study the effect of WIPBf over W. Based on these analyses, the

manager decided to slow down the activity not involving a higher number of W during 14th and 15

th

weeks. During these weeks, a larger WIPBf was deliberately generated maintaining a low W level. It

was determined that a WIPBf size closer to 2000 m2 could maximize labor productivity in order to

achieve PP levels of 800 m2 with W levels closer to 31 worker-weeks. During the 16th and 17

th

weeks the numbers of W was increased to take advantage of a higher WIPBf size.

A rough analysis of the data from Figure 7 shows that the mean actual PRI for the Predictions/no-

decisions period and the Predictions/decisions period is 70.55% and 100%, respectively. The effect

over labor productivity for the same period was estimated as the ratio between actual progress and

worker-weeks. Mean labor productivity for the Predictions/no-decisions period and

Predictions/decisions period was 20.4 (m2/worker-weeks) and 22.5 (m

2/worker-week), respectively.

In other words, planning reliability was increased by 41.0% and productivity by 10.3% (see Figure

7). A larger WIPBf size during weeks 14th and 15

th was produced, resulting in improved productivity

for the following weeks. The mean actual PRI for the 3rd

- 15th weeks is 72.49% and for the 16

th - 17

th

weeks is 100%, improving planning reliability by 38.0%. Similarly, the mean labor productivity for

24

the 3rd

- 15th weeks is 20.6 (m

2/worker-weeks) and for the 16

th - 17

th weeks is 27.0 (m

2/worker-weeks),

with a productivity improvement of 31.0%. The improvement in labor productivity is explained by a

better planning reliability and a direct action over production variables such as W and WIPBf using

the RCM-Plan as a decision-making aid tool.

6. Conclusions

This paper addressed the theoretical and practical framework that supports a computer prototype

called RCM-Plan based on Rational Commitment Model (RCM) principles, which allows to improve

reliability of operational work plans in construction projects. By using site production data such as

workers, buffers and plans, RCM-Plan automatically develops statistical models (multiple linear

regressions) to predict activity progress, and thus, support reliable work plans. One of the most

important features of the RCM-Plan is the explicit visualization in a multidimensional environment of

the interaction of several production parameters (workers, buffers and plans) and its impact on

planning predictions and labor capacity. This allows creating more reliable work plans, having in

mind a more realistic and a rational characterization of the production environment in projects. The

implementation of the RCM methodology can be complex but experimentation with RCM-Plan

allowed reducing implementation barriers in the organization of the case projects. Further research to

improve capabilities of RCM-Plan is necessary, which is part of the ongoing research of the authors.

References

Alarcón, L.F., Diethelm, S., Rojo, O. and Calderón R. (2005). Assessing the Impacts of Implementing

Lean Construction. Proceedings 13th Annual Conference of International Group for Lean

Construction, Sidney, Australia, July 19th – 21th.

Ballard, G. (2000). The Last Planner System of Production Control. PhD’s Dissertation, School of

Civil Engineering, Faculty of Engineering, The University of Birmingham, Birmingham, U.K.

Bustamante, J. A. (2007). Implementing of a Rational Commitment Model for Improving Planning

Reliability in Construction Projects. MEng’s Thesis, Faculty of Engineering, Pontificia Universidad

Católica de Chile, Santiago, Chile.

Choo, H.J. and Tommelein, I.D. (2000). WorkMovePlan: Database For Distributed Planning And

Coordination. Proceedings 8th Annual Conference of International Group for Lean Construction,

Brighton, UK, July 17th-19

th.

Choo, H.J., Tommelein, I.D., Ballard, G. and Zabelle, T.R. (1998). WorkPlan: Database for Work

Package Production Scheduling. Proceedings 6th Annual Conference of International Group for Lean

Construction, Guaruja, Brazil, August 13th-15

th.

25

Chua, D.K.H., Shen, L.J. and Bok, S.H. (1999). Integrated Production Scheduler for Construction

Look-Ahead Planning. Proceedings 7th Annual Conference of International Group for Lean

Construction, Berkeley, USA, July 26th-28

th.

González, V., Alarcón, L.F., Maturana, S., Mundaca F. and Bustamante, J. A. (2010). Improving

Planning Reliability and Project Performance using the Reliable Commitment Model. Journal of

Construction Engineering and Management, 10.1061/(ASCE)CO.1943-7862.0000215 (in-press).

González, V., Alarcón, L.F., Maturana, S., Mundaca, F., and Bustamante, J. A. (2009). Rational

Commitment Model: Improving Planning Reliability and Project Performance. Proceedings 17th

International Conference for Lean Construction, Taipei, Taiwan, July 15th – 17

th.

González, V., Alarcón, L.F. and Mundaca, F. (2008a). Investigating the Relationship between

Planning Reliability and Project Performance. Prod. Plan. Control, 19, (5), 461-474.

González, V., Alarcón. L.F., Maturana, S., Bustamante, J.A. and Mundaca, F. (2008b). Work-In-

Process Buffer Management using the Rational Commitment Model in Repetitive Projects.

Proceedings 16th Annual Conference of International Group for Lean Construction, Manchester,

UK, July 16th-18

th.

Huang, T., Kong, C.W., Guo, H.L., Baldwin, A. and Li, H. (2007). A Virtual Prototyping System for

Simulating Construction Processes. Autom. Constr., 16, (5), 576-585.

Hopp, W. J. and Spearman, M. L. (2000). Factory Physics: Foundations of Manufacturing

Management, 2nd

Ed., Irwin/McGraw-Hill, Boston.

Martínez, J. C. (1996). STROBOSCOPE: State and Resource Based Simulation of Construction

Processes. PhD’s Dissertation, University of Michigan, Ann Arbor, Mich.

P+C (2009). Project Plus Control. Accessed on December 12th, 2009.

See www.projectpluscontrol.cl/noticias.html

Sacks, R. and Harel, M. (2006). An Economic Game Theory Model of Subcontractor Resource

Allocation Behaviour. Const. Mgmt. and Econ., 24, (8), 869–881.

SPS|PM (2009). SPS|Production Manager. Accessed on December 12th, 2009.

See www.strategicprojectsolutions.com/Products_Services_SPSPM.aspx.

Sriprasert, E., Dawood, N. (2002). Next Generation of Construction Planning and Control System:

The Lewis Approach. Proceedings 16th Annual Conference of International Group for Lean

Construction, Gramado, Brazil, August 6th-8

th.

Tommelein, I. D., Riley, D. R. and Howell G. A. (1999). Parade Game: Impact of Work Flow

Variability on Trade Performance. J. Const. Engr. Mgmt., ASCE, 125, (5), 304-310.

26

http://www.projectpluscontrol.cl/noticias.htmlhttp://www.strategicprojectsolutions.com/Products_Services_SPSPM.aspx

Womack, J. P. and Jones, D. T. (1996). Lean Thinking: Banish Waste and Create Wealth in Your

Corporation, Simon & Schuster, New York.

27

Building Information Modelling in the Netherlands: A Status Report

van Nederveen, S.

Delft University of Technology

(email: g.a.vannederveen@tudelft.nl)

Beheshti, R.

Delft University of Technology

(email: m.r.beheshti@tudelft.nl)

Willems, P.

TNO

(email: peter.willems@tno.nl)

Abstract

Building Information Modelling (BIM) is nowadays widely accepted as a key enabler for innovation

in construction. In the Netherlands, people have been working on BIM for more than twenty years,

although most activities have been research efforts. But since the leading CAD vendors have

embraced BIM as a key development in CAD innovation, the implementation and use of BIM

technologies in practice have increased significantly.

Apart from various “pseudo-BIM”- initiatives (BIM-solutions within a single commercial software

platform, “closed” BIM-solutions that are not accessible by external parties, fancy CAD-solutions

presented as BIM-solutions), there are a number of interesting BIM-related developments in the

Netherlands.

The first development is the COINS project. This project aims for agreements for the storage and

exchange of construction objects. The main results of COINS are currently specifications for these

agreements and software tools for implementation of COINS-based systems. COINS uses the OWL

format for object definitions; interfaces with the IFC-models have also been developed. The COINS

project is initiated by the Dutch civil engineering industry, but the current focus is on products of the

entire building industry. Several pilot projects are currently taking place both in civil engineering and

in office and residential building.

The second development is somewhat different: it is the BIM Case Week, an initiative that brings

together professionals in the construction industry for a week, and lets them work together on a

design of a building project. The approach is fairly down-to-earth but it has provided very useful

insights in how exchange and sharing of construction information takes place in practice.

28

mailto:peter.willems@tno.nl

The third development is the Dynamic BIM initiative; this is currently an academic initiative that

aims at the support of project dynamics in a BIM context. The focus in this initiative is on innovative

design and engineering processes enabled by BIM technologies.

Keywords: building information modelling, the Netherlands, status report

29

1. Introduction

One of the current keywords in building innovation is Building Information Modelling, or BIM.

Since a few years, BIM is a buzzword. But Building Information Modelling is not a new activity or

technology; in fact people have been working on BIM (using different terms) for decades, although

mainly in research settings.

People in the Netherlands have been involved in BIM research since the early days of BIM, the

eighties of last century. Over the years, Dutch researchers have been working on BIM in various

projects. In recent years, a transition can be seen in the Netherlands from mainly research oriented

activities towards dissemination and implementation in building practice. More and more people and

companies become active with BIM. Articles on BIM appear in practice oriented journals, new

courses for building professionals are set up, and new dedicated BIM websites are set up, and so on.

This paper gives an overview of some important BIM-related developments in the Netherlands. First

a short historical overview of BIM in the Netherlands is given. Next the COINS project, the BIM

Case week and the Dynamic BIM initiative are discussed, followed by a short discussion of other

developments. But before all that, a short statement is made about the definition of BIM.

2. What is BIM?

There are a number of different definitions of BIM around. As more people are working with BIM,

this number increases, and as a consequence more misunderstandings occur. As long as BIM is

mainly a research topic, this is a little unpractical, but more or less unavoidable, just like with many

other definitions in research. But with the transition from BIM as a research topic towards BIM as a

commercial product or service, the need for a clear definition becomes really apparent. For example,

many companies claim they are doing BIM while critics say they are only offering smart CAD

solutions.

A useful definition for the term Building Information Modelling is the following by Lee et al (2006),

which is also used on Wikipedia: Building Information Modeling (BIM) is the process of

generating and managing building data during its life cycle. Typically it uses three-dimensional,

real-time, dynamic building modeling software to increase productivity in building design and

construction. The process produces the Building Information Model (also abbreviated BIM), which

encompasses building geometry, spatial relationships, geographic information, and quantities and

properties of building components

In addition, a useful definition for the term Building Information Model is the following by Van

Nederveen et al (2009): a Building Information Model is an information model of a building (or

building project) that comprises complete and sufficient information to support all lifecycle

processes, and which can be interpreted directly by computer applications. It comprises information

about the building itself as well as its components, and comprises information about properties such

as function, shape, material and processes for the building life cycle.

30

http://en.wikipedia.org/wiki/Building

The last definition is a little bit long term oriented, as life-cycle support by BIM is currently far from

common practice.

But let us go back to the initial question: what is BIM? A key question in this respect is: how can we

distinguish between “BIM” and “non-BIM”? For that purpose, the following characteristics of BIM

can be highlighted:

BIM aims at the exchange of semantic information. That is: the model that is developed does not only cover geometric information, but also material properties, functional information,

etc. For example, many advanced CAD systems that use concept of parametric modelling can

be very useful design aids. But if their internal model is solely based on geometric entities,

you cannot call these BIM modellers.

A prerequisite for BIM is the use of open standards. A Building Information Modeller may be a “closed” system, but the information that is exchanged or shared must be defined

according to an open standard, such as IFC. Although closed systems can be very effective, in

the long run they can lead to vendor-dependency and to outdated systems that are very

difficult to upgrade.

Neither of the definitions stated above explicitly mention open standards as a prerequisite for proper

BIM. Open standards are indeed often mentioned as a prerequisite. On the other hand, one can

question whether it is absolutely necessary to use for example IFCs in a BIM environment. In our

view this is an open issue.

3. History of BIM in the Netherlands

The Netherlands has quite a rich history in BIM research and development, which goes back to more

than twenty years ago. In the nineteen eighties, several groups in the Netherlands were involved in

research on CAD systems for architecture, and on the issue of data exchange between CAD systems.

The Dutch architectural CAD system Arcos/Arkey CAD was launched with some “building

intelligence” built in. A discussion started on the use of so-called reference models for CAD

exchange.

A key reference model in this context was the General AEC Reference Model by Wim Gielingh of

the Dutch research institute TNO (1988). This model was developed for the ISO STEP (ISO 10303)

project, and it provided a number of concepts and principles that we can regard now as BIM

concepts: as required and as designed information, generic-specific-occurrence information, life cycle

data, views on building data, etc. The famous “Hamburger” notation and the associated ideas are still

used in publications from all over the world.

Another interesting publication out of that period is the so-called IOP Bouw Informatie Model (Van

Merendonk and Van Dissel 1989). This model was the main end result of a large Dutch research

project aiming at the modelling of building information. Most of this publication consists of process

31

models in IDEF0, furthermore some data models have been presented in IDEFx. The models are

nowadays rarely used or referenced, but the title of the model is definitely remarkable.

In the early nineties, some very interesting BIM-related work was carried out in EU-projects in which

TNO was involved, such as ATLAS, PISA and COMBINE (Tolman 1999). In all of these projects,

product modelling based on ISO STEP played a key role. Many key concepts and principles of IFC

origin from these projects.

From the late nineties until today, a number of smaller scale national activities related to BIM took

place. Participants involved include among others the building specification organization STABU,

the organization for installation systems UNETO. Some of the initiatives have formed a platform

called PAIS, see www.paisbouw.nl. A significant national development has been VISI, a standard for

communication in building projects based on transactions and messages, see www.visi.nl. VISI uses

protocols for common communication processes using transactions that consist of a sequence of

messages between participants.

At the moment there are a number of interesting BIM developments going on in the Netherlands.

Three developments will be discussed in the next three sections of this paper. The first development

is the COINS project. This project is interesting because many key players from the Dutch

construction industry are participating. The approach taken can be regarded as pragmatic, yet they do

use an open standards approach based on IFC and OWL.

The second development is the BIM Case Week. This initiative brings together professionals in the

construction industry for a week, and lets them work together on a design of a building project. The

BIM Case Week is similar to the Build London Live events in the UK. Its biggest value is the great

amount of public attention for BIM that it attracts.

The third development is the Dynamic BIM initiative. This is currently an academic initiative that

aims at the support of project dynamics in a BIM context. This initiative is particularly interesting

because it tries to bring BIM another step further through new research and innovation.

An important development in the Netherlands that is not directly about BIM, but that has a significant

impact on BIM work, has been the growing interest in Systems Engineering. Since the late nineties,

Systems Engineering was introduced at the large infrastructure principals ProRail and

Rijkswaterstaat, when these organizations became involved in large scale projects such as the High

Speed Link railway project between Amsterdam and Paris. With Systems Engineering, infrastructure

projects became more formal, with explicit procedures for requirements management, verification &

validation and risk management. Of course the companies that work for Rijkswaterstaat and ProRail

had to follow the Systems Engineering process, which meant that almost the entire civil engineering

sector had to deal with Systems Engineering. The impact of Systems Engineering on BIM work in the

Netherlands can be seen in current developments that will be discussed below.

32

http://www.paisbouw.nl/http://www.visi.nl/

4. The COINS project

The first Dutch development to be discussed in this paper is the COINS project. This project aims for

agreements for the storage and exchange of construction objects. The acronym COINS stands for

„Construction Objects and the INtegration of processes and Systems (see www.coinsweb.nl and click

on “Introduction COINS program”).

The COINS project was started in 2003 by a number of organizations from